AI/ML Engineer for Grounded Scientific Automation

Breakout Ventures; Cambridge (MA); On-site; USD 150,000 - 200,000; Full time

Benefits offered by this job

Competitive compensation

Full benefits

Job summary

Transfyr seeks an AI/ML Engineer to build end-to-end ML systems that learn from messy laboratory data. You will collaborate with perception engineers to ground data into actionable interpretations and integrate models into real workflows, shaping automation in science.

You will work on models that handle partial observability, noise, and changing protocols, while communicating uncertainty and justification to scientists and operators.

Qualifications

• Strong programming experience in software engineering, data systems, and AI/ML product development.
• Solid grounding in machine learning with real-world data learning capabilities.
• Experience reasoning about multimodal data (vision, audio, sensor data) and how signals interact.

Responsibilities

• Build end-to-end ML systems that learn from messy, real-world data captured in active lab environments.
• Collaborate with perception engineers to design multimodal pipelines and grounded interpretations of actions and outcomes.
• Develop models that reason about why experiments succeed or fail when multiple factors are intertwined.
• Create systems that indicate model confidence to guide human intervention when needed.
• Integrate models into real workflows where outputs influence human decisions and robotic actions.
• Ensure models generalize across experiments, teams, and geographies, not just lab-specific artifacts.

Skills

Great programmer

ML Fundamentals

Multimodal Learning

Python & Frameworks

Data & Pipelines

Production Awareness

Systems Mindset

Learning Velocity

Tools

Python

PyTorch

Job description

Transfyr seeks an AI/ML Engineer to build end-to-end ML systems that learn from messy laboratory data. You will collaborate with perception engineers to ground data into actionable interpretations and integrate models into real workflows, shaping automation in science.

You will work on models that handle partial observability, noise, and changing protocols, while communicating uncertainty and justification to scientists and operators.